In general, the prefetching of HTTP content is intended to reduce the latency in downloading web pages that include references to multiple objects by issuing requests from a proxy server and pushing the responses to a proxy client without waiting for the browser to request the items. The proxy server identifies objects to be prefetched by scanning the content of HTML, CSS, JavaScript™, or other objects to identify references to items that are likely to be requested by the browser when it receives the object containing the reference. The browser will not request items that are in its cache if the objects are fresh under the HTTP caching rules. Also, if the browser has an item which may be stale, it will issue a request with the “If-Modified-Since” directive, so that if the object has not changed, a short HTTP response can be returned instead of the full file.
Thus, without a model of the browser cache, the prefetching proxy server would send many objects that will not be requested by the browser and send full files where the browser would normally have received the “Not-Modified” responses. This creates extra load on both the proxy server to webhost upstream link and the proxy server to the client accelerated link. Thus, the current system downloads a number of unnecessary objects, which in turn causes slowdown of the link. Hence, improvements in the art are needed.